Some Questions about Elastic Tensor Calculations in VASP: Dimensionality Applicability and Experimental Deviation.

[hszhao.cn@gmail.com]

Dear VASP Forum,

I have several questions regarding elastic tensor calculations using IBRION=5,6 in combination with ISIF=3:

1. Is this calculation method applicable to 1D, 2D, and 3D structures? If not, what are the limitations?
2. Is it possible to derive 1D and 2D elastic tensors from 3D elastic tensor calculations?
3. What is the typical order of magnitude of deviation between these calculated results and experimental measurements?

Thank you for your assistance.

Best regards,
Zhao

[ahampel]

Hi Zhao,

It should be possible without problems to perform elastic tensor calculations for 1D, 2D, or 3D materials. As long as you follow the usual guideline that those systems are not truly 1D or 2D, but with a vacuum in the "empty" directions. Just make sure to have the vacuum direction for example in 2D to be clearly separated from the other lattice directions, i.e. only a z-component and orthogonal to the other two unit vectors a and b. Then you can simply ignore the output for the elastic tensor for the third direction. There might be of course some small overhead since VASP still will apply small stresses in these directions but the result should be correct. In the same way it should be possible to derive then 1D and 2D elastic tensors if I understand you correctly? But VASP does not allow you in this mode to explicitly strain only 2D let's say. This you would need to do manually. See also https://mattermodeling.stackexchange.co ... -practices for details on how to do this with an external tool.

For the third question I cannot really help you. I think this is a research question / literature.

Best,
Alex

[hszhao.cn@gmail.com]

Dear Alex,

In the same way it should be possible to derive then 1D and 2D elastic tensors if I understand you correctly?

I mean, can we derive the corresponding elastic tensor for a crystal plane, e.g., (111) or crystal direction, e.g., [123], from the computed 3D elastic tensor?

Regards,
Zhao

[ahampel]

Hi Zhao,

I see that you basically asked this question already back here: https://vasp.at/forum/viewtopic.php?t=19829 . The answer is basically: "If you compute the elastic tensor in the primitive box you can use symmetry operations to transform the elastic tensor to the symmetry of the conventional box." And this paper discusses how to do this: https://hal.science/hal-02410330/document . VASP does not have any functionality to perform these transformations but in general this should be covered in the calculation of the elastic tensor. Note, this changes of course if you break the symmetry of your system, e.g. if you explicitly construct a (111) surface (with vacuum) the result will be different and the calculation should be performed for this POSCAR.

Best,
Alex